ELECTRIC GRIDAI and the Future of the U.S. Electric Grid

By Doug Irving

Published 11 April 2025

Despite its age, the U.S. electric grid remains one of the great workhorses of modern life. Whether it can maintain that performance over the next few years may determine how well the U.S. competes in an AI-driven world.

Despite its age, the U.S. electric grid remains one of the great workhorses of modern life. Whether it can maintain that performance over the next few years may determine how well the U.S. competes in an AI-driven world.

AI is a big part of the challenge. Its vast data centers suck up energy like small cities. But a recent RAND study suggests AI could be a big part of the solution, too. There are risks here—some obvious, some not—and grid operators need to move with caution. But AI could usher in an energy future that is more resilient, more efficient, and more affordable for customers.

“The important thing is to do it without rocking the boat,” said Ismael Arciniegas Rueda, a former energy executive, now a senior economist at RAND. “The grid can fail, definitely, and I don’t think people understand the consequences if that does happen. It’s not just the lights going out. Our whole life depends on whether or not energy is available 100 percent of the time.”

Companies working with AI have warned that they are already struggling to find the power they need. Keeping them on U.S. soil has become a national imperative, especially in light of the deepening competition with China. That means upgrading and modernizing the grid, much of which was built in the 1960s and ‘70s.

Grid operators in the U.S. and globally have already started to use AI to monitor transmission lines and isolate faults. AI systems are also analyzing huge amounts of data in real-time to better predict fluctuations in supply and demand. They’re creating a grid that operates by the second, much faster than human operators can respond.

Arciniegas Rueda wanted to know whether the benefits really pencil out—for customers as well as for companies. He teamed up with researchers from RAND Europe to look at two AI applications that have started to come online in recent years. One acts as a kind of cruise control for energy use in commercial and industrial buildings. It automatically adjusts their heating and cooling systems to maintain temperatures more efficiently. The other smooths out demand spikes, shifting energy use to lower-cost times of day. It might, for example, wait until evening to switch on a washing machine.